[ENH] Feature as Predictor

Orange/classification/tests/test_column.py

@@ -0,0 +1,165 @@
+import unittest
+from unittest.mock import patch
+
+import numpy as np
+
+from Orange.classification import ColumnLearner, ColumnClassifier
+from Orange.data import DiscreteVariable, ContinuousVariable, Domain, Table
+
+
+class ColumnTest(unittest.TestCase):
+    @classmethod
+    def setUpClass(cls):
+        cls.domain = Domain([DiscreteVariable("d1", values=["a", "b"]),
+                             DiscreteVariable("d2", values=["c", "d"]),
+                             DiscreteVariable("d3", values=["d", "c"]),
+                             ContinuousVariable("c1"),
+                             ContinuousVariable("c2")
+                            ],
+                            DiscreteVariable("cls", values=["c", "d"]),
+                            [DiscreteVariable("m1", values=["a", "b"]),
+                             DiscreteVariable("m2", values=["d"]),
+                             ContinuousVariable("c3")]
+                            )
+        cls.data = Table.from_numpy(
+            cls.domain,
+            np.array([[0, 0, 0, 1, 0.5],
+                      [0, 1, 1, 0.25, -3],
+                      [1, 0, np.nan, np.nan, np.nan]]),
+            np.array([0, 1, 1]),
+            np.array([[0, 0, 2],
+                      [1, 0, 8],
+                      [np.nan, np.nan, 5]])
+        )
+
+    @patch("Orange.classification.column.ColumnModel")
+    def test_fit_storage(self, clsfr):
+        learner = ColumnLearner(self.domain.class_var, self.domain["d2"])
+        self.assertEqual(learner.name, "column 'd2'")
+        learner.fit_storage(self.data)
+        clsfr.assert_called_with(self.domain.class_var, self.domain["d2"], None, None)
+
+        learner = ColumnLearner(self.domain.class_var, self.domain["c3"])
+        learner.fit_storage(self.data)
+        clsfr.assert_called_with(self.domain.class_var, self.domain["c3"], None, None)
+
+        learner = ColumnLearner(self.domain.class_var, self.domain["c3"], 42, 3.5)
+        self.assertEqual(learner.name, "column 'c3'")
+        learner.fit_storage(self.data)
+        clsfr.assert_called_with(self.domain.class_var, self.domain["c3"], 42, 3.5)
+
+    def test_classifier_init_checks(self):
+        cls = ColumnClassifier(self.domain.class_var, self.domain["d2"])
+        cls.name = "column 'd2'"
+
+        cls = ColumnClassifier(self.domain.class_var, self.domain["d3"])
+        cls.name = "column 'd3'"
+
+        cls = ColumnClassifier(self.domain.class_var, self.domain["c3"])
+        cls.name = "column 'c3'"
+
+        self.assertRaises(
+            ValueError,
+            ColumnClassifier,
+            self.domain.class_var, self.domain["d1"])
+
+        self.assertRaises(
+            ValueError,
+            ColumnClassifier,
+            DiscreteVariable("x", values=("a", "b", "c")), self.domain["c3"])
+
+    def test_check_prob_range(self):
+        self.assertTrue(
+            ColumnClassifier.valid_prob_range(np.array([0, 0.5, 1]))
+        )
+        self.assertTrue(
+            ColumnClassifier.valid_prob_range(np.array([0, 0.5, np.nan]))
+        )
+        self.assertFalse(
+            ColumnClassifier.valid_prob_range(np.array([0, 0.5, 1.5]))
+        )
+        self.assertFalse(
+            ColumnClassifier.valid_prob_range(np.array([0, 0.5, -1]))
+        )
+
+    def test_check_value_sets(self):
+        d1, d2, d3, *_ = self.domain.attributes
+        c = self.domain.class_var
+        m2: DiscreteVariable = self.domain["m2"]
+        self.assertFalse(ColumnClassifier.valid_value_sets(c, d1))
+        self.assertTrue(ColumnClassifier.valid_value_sets(c, d2))
+        self.assertTrue(ColumnClassifier.valid_value_sets(c, d3))
+        self.assertTrue(ColumnClassifier.valid_value_sets(c, m2))
+        self.assertFalse(ColumnClassifier.valid_value_sets(m2, c))
+
+    def test_predict_discrete(self):
+        # Just copy
+        model = ColumnClassifier(self.domain.class_var, self.domain["d2"])
+        self.assertEqual(model.name, "column 'd2'")
+        classes, probs = model(self.data, model.ValueProbs)
+        np.testing.assert_equal(classes, [0, 1, 0])
+        np.testing.assert_equal(probs, [[1, 0], [0, 1], [1, 0]])
+
+        # Values are not in the same order -> map
+        model = ColumnClassifier(self.domain.class_var, self.domain["d3"])
+        classes, probs = model(self.data, model.ValueProbs)
+        np.testing.assert_equal(classes, [1, 0, np.nan])
+        np.testing.assert_equal(probs, [[0, 1], [1, 0], [0.5, 0.5]])
+
+        # Not in the same order, and one is missing -> map
+        model = ColumnClassifier(self.domain.class_var, self.domain["m2"])
+        classes, probs = model(self.data, model.ValueProbs)
+        np.testing.assert_equal(classes, [1, 1, np.nan])
+        np.testing.assert_equal(probs, [[0, 1], [0, 1], [0.5, 0.5]])
+
+        # Non-binary class
+        domain = Domain(
+            self.domain.attributes,
+            DiscreteVariable("cls", values=["a", "c", "b", "d", "e"]))
+        data = Table.from_numpy(domain, self.data.X, self.data.Y)
+        model = ColumnClassifier(domain.class_var, domain["d3"])
+        classes, probs = model(data, model.ValueProbs)
+        np.testing.assert_equal(classes, [3, 1, np.nan])
+        np.testing.assert_almost_equal(
+            probs,
+            np.array([[0, 0, 0, 1, 0],
+                      [0, 1, 0, 0, 0],
+                      [0.2, 0.2, 0.2, 0.2, 0.2]]))
+
+    def test_predict_as_direct_probs(self):
+        model = ColumnClassifier(self.domain.class_var, self.domain["c1"])
+        self.assertEqual(model.name, "column 'c1'")
+        classes, probs = model(self.data, model.ValueProbs)
+        np.testing.assert_equal(classes, [1, 0, np.nan])
+        np.testing.assert_equal(probs, [[0, 1], [0.75, 0.25], [0.5, 0.5]])
+
+        model = ColumnClassifier(self.domain.class_var, self.domain["c2"])
+        self.assertRaises(ValueError, model, self.data)
+
+        model = ColumnClassifier(self.domain.class_var, self.domain["c3"])
+        self.assertRaises(ValueError, model, self.data)
+
+    def test_predict_with_logistic(self):
+        model = ColumnClassifier(
+            self.domain.class_var, self.domain["c1"], 0.5, 3)
+        classes, probs = model(self.data, model.ValueProbs)
+        np.testing.assert_equal(classes, [1, 0, np.nan])
+        np.testing.assert_almost_equal(
+            probs[:, 1], [1 / (1 + np.exp(-3 * (1 - 0.5))),
+                          1 / (1 + np.exp(-3 * (0.25 - 0.5))),
+                          0.5])
+        np.testing.assert_equal(probs[:, 0], 1 - probs[:, 1])
+
+        model = ColumnClassifier(
+            self.domain.class_var, self.domain["c2"], 0.5, 3)
+        classes, probs = model(self.data, model.ValueProbs)
+        np.testing.assert_equal(classes, [0, 0, np.nan])
+        np.testing.assert_almost_equal(
+            probs[:, 1], [1 / (1 + np.exp(-3 * (0.5 - 0.5))),
+                          1 / (1 + np.exp(-3 * (-3 - 0.5))),
+                          0.5])
+        np.testing.assert_equal(probs[:, 0], 1 - probs[:, 1])
+
+
+if __name__ == "__main__":
+    unittest.main()

Orange/modelling/__init__.py

@@ -11,6 +11,7 @@
 from .randomforest import *
 from .svm import *
 from .tree import *
+from .column import *
 try:
     from .catgb import *
 except ImportError:

Orange/modelling/column.py

@@ -0,0 +1,155 @@
+from typing import Optional
+
+import numpy as np
+
+from Orange.data import Variable, DiscreteVariable, Domain, Table
+from Orange.classification import LogisticRegressionLearner
+from Orange.regression import LinearRegressionLearner
+from Orange.modelling import Model, Learner
+
+__all__ = ["ColumnLearner", "ColumnModel"]
+
+
+def _check_column_combinations(
+        class_var: Variable,
+        column: Variable,
+        fit_regression: bool):
+    if class_var.is_continuous:
+        if not column.is_continuous:
+            raise ValueError(
+                "Regression can only be used with numeric variables")
+        return
+
+    assert isinstance(class_var, DiscreteVariable)  # remove type warnings
+    if column.is_continuous:
+        if len(class_var.values) != 2:
+            raise ValueError(
+                "Numeric columns can only be used with binary class variables")
+    else:
+        assert isinstance(column, DiscreteVariable)
+        if not valid_value_sets(class_var, column):
+            raise ValueError(
+                "Column contains values that are not in class variable")
+    if fit_regression and not column.is_continuous:
+        raise ValueError(
+            "Intercept and coefficient are only allowed for continuous "
+            "variables")
+
+
+def valid_prob_range(values: np.ndarray):
+    return np.nanmin(values) >= 0 and np.nanmax(values) <= 1
+
+
+def valid_value_sets(class_var: DiscreteVariable,
+                     column_var: DiscreteVariable):
+    return set(column_var.values) <= set(class_var.values)
+
+
+class ColumnLearner(Learner):
+    def __init__(self,
+                 class_var: Variable,
+                 column: Variable,
+                 fit_regression: bool = False):
+        super().__init__()
+        _check_column_combinations(class_var, column, fit_regression)
+        self.class_var = class_var
+        self.column = column
+        self.fit_regression = fit_regression
+        self.name = f"column '{column.name}'"
+
+    def __fit_coefficients(self, data: Table):
+        # Use learners from Orange rather than directly calling
+        # scikit-learn, so that we make sure we use the same parameters
+        # and get the same result as we would if we used the widgets.
+        data1 = data.transform(Domain([self.column], self.class_var))
+        if self.class_var.is_discrete:
+            model = LogisticRegressionLearner()(data1)
+            return model.intercept[0], model.coefficients[0][0]
+        else:
+            model = LinearRegressionLearner()(data1)
+            return model.intercept, model.coefficients[0]
+
+    def fit_storage(self, data: Table):
+        if data.domain.class_var != self.class_var:
+            raise ValueError("Class variable does not match the data")
+        if not self.fit_regression:
+            return ColumnModel(self.class_var, self.column)
+
+        intercept, coefficient = self.__fit_coefficients(data)
+        return ColumnModel(self.class_var, self.column, intercept, coefficient)
+
+
+class ColumnModel(Model):
+    def __init__(self,
+                 class_var: Variable,
+                 column: Variable,
+                 intercept: Optional[float] = None,
+                 coefficient: Optional[float] = None):
+        super().__init__(Domain([column], class_var))
+
+        _check_column_combinations(class_var, column, intercept is not None)
+        if (intercept is not None) is not (coefficient is not None):
+            raise ValueError(
+                "Intercept and coefficient must both be provided or absent")
+
+        self.class_var = class_var
+        self.column = column
+        self.intercept = intercept
+        self.coefficient = coefficient
+        if (column.is_discrete and
+                class_var.values[:len(column.values)] != column.values):
+            self.value_mapping = np.array([class_var.to_val(x)
+                                          for x in column.values])
+        else:
+            self.value_mapping = None
+
+        pars = f" ({intercept}, {coefficient})" if intercept is not None else ""
+        self.name = f"column '{column.name}'{pars}"
+
+    def predict_storage(self, data: Table):
+        vals = data.get_column(self.column)
+        if self.class_var.is_discrete:
+            return self._predict_discrete(vals)
+        else:
+            return self._predict_continuous(vals)
+
+    def _predict_discrete(self, vals):
+        assert isinstance(self.class_var, DiscreteVariable)
+        nclasses = len(self.class_var.values)
+        proba = np.full((len(vals), nclasses), np.nan)
+        rows = np.isfinite(vals)
+        if self.column.is_discrete:
+            mapped = vals[rows].astype(int)
+            if self.value_mapping is not None:
+                mapped = self.value_mapping[mapped]
+                vals = vals.copy()
+                vals[rows] = mapped
+            proba[rows] = 0
+            proba[rows, mapped] = 1
+        else:
+            if self.coefficient is None:
+                if not valid_prob_range(vals):
+                    raise ValueError("Column values must be in [0, 1] range "
+                                     "unless logistic function is applied")
+                proba[rows, 1] = vals[rows]
+            else:
+                proba[rows, 1] = (
+                    1 /
+                    (1 + np.exp(-self.intercept - self.coefficient * vals[rows])
+                     ))
+
+            proba[rows, 0] = 1 - proba[rows, 1]
+            vals = (proba[:, 1] > 0.5).astype(float)
+            vals[~rows] = np.nan
+        return vals, proba
+
+    def _predict_continuous(self, vals):
+        if self.coefficient is None:
+            return vals
+        else:
+            return vals * self.coefficient + self.intercept
+
+    def __str__(self):
+        pars = f" ({self.intercept}, {self.coefficient})" \
+            if self.intercept is not None else ""
+        return f'ColumnModel {self.column.name}{pars}'